<<

Clay (1999) 34, 7–25

The origin and formation of clay minerals in soils: past, present and future perspectives

M. J. WILSON

Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen AB15 8QH, UK

(Received 23 September 1997; revised 15 January 1998)

ABSTRACT: The origin and formation of soil clay minerals, namely , vermiculites, smectites, chlorites and interlayered minerals, interstratified minerals and kaolin minerals, are broadly reviewed in the context of research over the past half century. In particular, the pioneer overviews of Millot, Pedro and Duchaufour in France and of Jackson in the USA, are considered in the light of selected examples from the huge volume of work that has since taken place on this topic. It is concluded that these early overviews may still be regarded as being generally valid, although it may be that too much emphasis has been placed upon transformation mechanisms and not enough upon neoformation processes. This review also highlights some of the many problems pertaining to the origin and formation of soil clays that remain to be resolved.

It has long been recognized that the minerals in the detail remained to be filled in, as well as a time that clay (<2 mm) fractions of soils play a crucial role in immediately pre-dated the widespread utilization in determining their major physical and chemical soil science of analytical techniques such as properties, and inevitably, questions concerning scanning electron microscopy, electron probe the origin and formation of these minerals have microanalysis, Mo¨ssbauer spectroscopy, electron assumed some prominence in soil science research. spin resonance spectroscopy and infrared spectro- This review considers some important aspects of scopy. This review will therefore attempt to these questions and is confined to clay minerals in summarize the general conclusions that had been soils as they are generally understood, that is as a arrived at regarding the origin and formation of medium for plant growth. certain clay minerals in soils in the early 1960s, the The occurrence of clay minerals in saprolites or developments since that time and the situation as it weathered is not discussed to any great extent currently stands, and to consider the outstanding because many studies have shown that saprolites are problems to be addressed in the future. In this characterized by physicochemical conditions that context, particular attention will be paid to micas, are different to those in soils. Intimate grain-to- vermiculites, smectites, chlorites, interstratified grain contacts promote a special chemical environ- minerals and kaolin minerals. As in all such ment on a local scale, bringing about the formation reviews, a certain amount of subjectivity and of transient clay phases which quickly personal preference is perhaps inevitable. disappear in the overlying soil. For the purposes of this review the perspective of PAST PERSPECTIVES the past will be taken to be that of the early 1960s, which is a convenient historical benchmark in that The reviews of Millot (1965) and Jackson (1964) it represents a time when the overall structural represent a convenient starting point with regard to characteristics of the layer silicate clay minerals the origin and formation of soil clays as assessed were to a large extent known, even though much principally by the X-ray diffraction and microscopic

# 1999 The Mineralogical Society 8 M. J. Wilson techniques that were available at the time. Millot drained, acid and -depleted tropical environ- distinguished three principal processes to account ments, where an abundant supply of water ensures for the genesis of clay minerals, which may occur the required silica and alumina. is at different points in the geochemical cycle more typical of a poorly drained or hydromorphic including or soil formation at the soils under alkaline conditions, rich in Mg and Ca earth’s surface. These processes are: (a) detrital and where Si, Al and commonly Fe tend to inheritance whereby, for soils, clay minerals are accumulate. inherited from pre-existing parent rock or weathered These generalizations were further developed and materials; (b) transformation where the essential synthesized by French soil scientists, most notably silicate structure of the is maintained Pedro and Duchaufour. Pedro (1964) distinguished to a large extent, but with major change in the two mineral weathering processes implicated in the interlayer region of the structure; and (c) neoforma- formation of clay minerals in soils which he termed tion, where the clay mineral forms through crystal- ‘hydrolysis’ and ‘acidolysis’. (A summary of this lization of gels or solutions. work in English is given in Pedro (1982)). Inherited soil clays may be of an extremely Hydrolysis of minerals occurs through dilute diverse and complex nature, reflecting both the solutions in the pH range 5À9.6 and may be total variety of the parent rock as well as the or partial. Total hydrolysis leads to the removal of transformation and neoformation processes that all elements including silica, and to the precipita- may have occurred in previous weathering environ- tion of gibbsite and minerals, whereas ments. In order to understand current clay mineral- partial hydrolysis, under different conditions, leads forming processes in a soil, it is essential that the to the formation of smectite minerals (Pedro, 1982). contribution of inheritance is clearly understood. Acidolysis operates when the soil solution pH is <5 With regard to transformation of clay minerals, a or has strongly complexing properties and can again general example would be: be total or limited. Total acidolysis involves complete solubilization of minerals with no ? ? smectite. precipitation of Al. Limited acidolysis leads to the This reaction proceeds through a process of fixation of Al in octahedral and interlamellar depletion and exchange of interlayer K and positions in layer silicates. In general, acidolysis concomitant decrease of layer charge. Such is associated with podzols, podzolic brown soils and changes are, however, deceptively simple and acid brown soils of cold temperate climates, have given rise to much debate about the precise whereas hydrolysis is dominant in ferrallitic soils mechanisms involved, as will be discussed later. of the warm humid tropics and is prominent in Millot distinguished ‘degradation’ and ‘aggradation’ warm temperate zones and in the dry subtropics as separate forms of the transformation process. The (Table 1). above conversion of illite to montmorillonite A similar synthesis relating to soil clays and involves depletion of elements from illite and is weathering was described by Duchaufour (1960) termed degradation, but the reverse reaction who distinguished geochemical weathering under (aggradation) involves addition of K and other near neutral conditions with no organic acid anions elements. Millot considered that degradation was and typical of tropical environments, and biochem- characteristic of weathering rocks and soils, but that ical weathering under acid conditions with organic aggradation was rare in such environments. anions and typical of temperate climates. The The formation of clay minerals through neofor- former is characterized by neoformed clay minerals mation clearly depends upon the appropriate and the latter by clay minerals formed by physicochemical conditions of the immediate transformation. weathering environment, such as the pH, composi- The overview of Jackson (1964) concerning the tion and concentration of the soil solutions, as well distribution, stability and weathering reactions of as the nature of the starting material and factors clay minerals in soils is, largely, consistent with the relating to the external environment like tempera- conclusions of the French soil scientists. Jackson ture, rainfall and percolation rate. Millot described described the dominant clay mineral types occur- kaolinite and montmorillonite as classical products ring in the Soil Orders of the taxonomy of the of neoformation in soils forming under contrasting USDA Soil Survey Staff (1960). There was a conditions. Thus, kaolinite is typical of freely predominance of micaceous, interstratified or Origin and formation of clay minerals in soils 9

TABLE 1. Occurrence of soil clay minerals in relation to weathering process, principal mechanism and soil types (Adapted from Pedro, 1982).

Weathering Principal Clay Principal process mechanism minerals soil types

Transformation Acidolysis Smectites Spodosols (partial) Vermiculites Inceptisols Al-intergrades Entisols Al-chlorite Neoformation Hydrolysis Smectites Spodosols (partial) Mollisols Alfisols

Hydrolysis Kaolinite Ultisols (total) Oxisols Gibbsite

interlayered layer silicate minerals in soils such as 2:1?1:1 pedogeochemical reactions. A major role Entisols, Inceptisols and Spodosols which occur for precursor crystals was also suggested in the widely in temperate climates. Clay minerals in soils growth of montmorillonite with hydroxy units of of tropical climates such as Oxisols and Ultisols, Al, Fe and Mg forming on pre-existing mont- were dominated by kaolinite and halloysite, in morillonite crystallites, which become silicated to addition to gibbsite and sesquioxides. However, the form new montmorillonite layers. Jackson (1964) occurrence of 2:1 to 2:2 intergrades was also concluded that ‘‘the almost universal occurrence of indicated in both Oxisols and Ultisols and in the hydroxy cation units precipitated on the surfaces of latter, vermiculite was also mentioned. In the layer silicates exposed to weathering is thus weathering reactions described by Jackson the extremely fundamental to clay mineralogy and emphasis was placed very much upon transforma- clay genesis’’. tion pathways (Fig. 1) and particularly on the role of hydroxy Al interlayers. Thus, it was envisaged PRESENT PERSPECTIVES that both smectite and vermiculite minerals could accumulate interlayers during extensive This section is concerned with research bearing on weathering, to the point where they approach a the origin of clay minerals in soils, in a global mineral resembling dioctahedral chlorite. A subse- context, which post-dates the seminal French and quent weathering step involving tetrahedra inver- American work described above. In particular, it is sion could then give kaolinite (or halloysite), of interest to consider whether the major conclu- although kaolinite could also form from sions reached need to be modified, refined or even directly. Kaolinite formation was thus proposed as abandoned in the light of the enormously increased proceeding through 2:1?2:2?1:1 or through volume of data now available on clay minerals in

FIG. 1. Pathways for the formation of soil clay minerals as outlined by Jackson (1964). 10 M. J. Wilson soils, often gained through the utilization of new these Hawaiian soils. However, convincing techniques. Comprehensive reviews of the literature evidence for neoformed mica in some Australian on soil clay minerals up until 1989 are available in soils was presented by Norrish & Pickering (1983). the relevant chapters of the monograph entitled A neoformational origin was indicated by the Minerals in Soil Environments (Dixon & Weed, perfect platy hexagonal morphology of the clay 1989) published by the Soil Science Society of mica and it is significant that the mineral is Fe-rich, America. This forms a very useful starting point for in this respect resembling which forms the present review where the emphasis will be in surface sediments under marine conditions. It is placed upon recent work. not clear, however, that the mica described by Norrish & Pickering (1983) is truly pedogenic in that the soil was derived from lacustrine sediments. Mica minerals It is possible that the mica may have formed in It is now clear, as it has been for some special conditions obtained during the drying out of considerable time (Fanning et al., 1989), that the the lake bed, rather than during soil formation on micaceous minerals that occur in soil clay fractions the lake bed sediments. have been inherited largely from parent rock or A general point made by Fanning et al. (1989) other material, where they originally formed under was that, at that time, there were very few detailed different PÀT conditions from those pertaining at studies of the crystallochemistry of soil clay micas. the earth’s surface. However, there is some This is true even now but there have been some evidence that mica may form pedogenically, but interesting advances in the nineties brought about only in special circumstances. Thus, Niederbudde by the use of analytical electron microscopy, (1975) and Niederbudde & Kussmaul (1978) particularly by Robert and co-workers in presented evidence to support the formation of Versailles. Although it may be agreed that virtually mica-like clay minerals in fertilized loess-derived all soil micas are of inherited origin, it is still soils in Germany following K fixation by beidellitic important to determine what changes may have smectites. The conversion of K-saturated smectites occurred to the clay size micas under pedogenetic to mica-like products following extensive wetting conditions, bearing in mind the importance and drying cycles in the laboratory, and involving attributed to micas as precursors of interstratified, change from turbostratic stacking to a semi-ordered expansible and interlayered minerals in soils. structure (de la Calle & Suquet, 1988), suggests a Robert et al. (1991) using HRTEM on impregnated possible mechanism for such a change. Similarly, it sections of French soils derived from sediments, has been suggested that the clay micas found in the showed that many of the particles were exceedingly surface horizons of soils of arid environments, such thin, recalling the fundamental illite particles as south west USA (Nettleton et al., 1973) and Iran described by Nadeau et al. (1984) in their studies (Mahjoory, 1975) are pedogenic in origin. In both of interstratified clays. Similar findings for clay-size instances a transformation process was envisaged micas in Iowa soils were arrived at independently whereby it was possible to fix K in pre-existing by Laird & Nater (1993). Robert et al. (1991) went smectites because of the hot and dry soil conditions. on to distinguish two types of mica mineral in the The need for caution in proposing mica formation clay fractions of the soils they studied (Table 2). by pedogenesis is emphasized by previously The first, described as micromica, occurs in mistaken interpretations concerning concentrations of clay mica in the surface horizons of Hawaiian TABLE 2. Nature of illitic clays in soils (after Robert et soils. Initially, it was thought that this mica must be al., 1991). pedogenic because mica was undetectable in the sub-surface soil and the basaltic parent materials Characteristics Micromica Illite (sensu stricto) (Juang & Uehara, 1968) but later work using isotopic analysis showed conclusively that the abplane micrometric nanometric mica mineral was older than the basalt flow from Thickness >10 nm ~5 nm which the soil was derived and therefore could not K2O ~10% ~7.5% be pedogenic (Dymond et al., 1974). Aeolian Charge >0.9 <0.6 deposition following tropospheric transport almost Structure 3-dimensional 3-dimensional certainly explains the origin of the mica minerals in Origin and formation of clay minerals in soils 11 relatively thick particles of micrometric size with a conceivable that oxidized but otherwise relatively full content of K and is highly charged. The second unaltered biotite fragments may eventually find type is termed illite (sensu stricto) and occurs in their way into soil clay fractions. thin particles of nanometric size, is somewhat depleted in K and is less highly charged. These Vermiculite minerals two types of micaceous clays can exist in composite crystals, implying that there is a genetic relationship The review of Douglas (1989) concerning the between them. Later work on the weathering of origin and formation of vermiculite minerals in micaceous clays in soils (Romero et al., 1992; soils emphasized a number of significant points. Aouidjit et al., 1996) does indeed show that particle Firstly, there is general agreement that most microdivision is an important part of the process. vermiculites are considered to form by the weath- Recent studies of Spanish soils developed upon ering of mica as was illustrated early on by the highly micaceous parent materials presented work of Walker (1949). It is true that there is some chemical and structural evidence for the conversion evidence to support the formation of vermiculite by of 2M1 muscovitic mica to the 1Md polytype with the weathering of chlorite, involving the preferential decreasing grain size (Martı´n Garcı´a et al., 1997). decomposition of the interlayer hydroxide sheet The picture that is emerging from these investiga- (Stephen, 1952; Ross & Kodama, 1976; Makumbi tions is that illitic clays may evolve in soil from the & Herbillon, 1972), or even through the decom- weathering of dioctahedral muscovitic mica position of non layer-silicate minerals such as minerals involving loss of K and layer charge and orthopyroxene (Basham, 1974), but in most principally mediated by the physical process of instances there seems no need to look further than particle microdivision. a genetic relationship with a precursor mica. This Although the vast majority of micaceous soil link is particularly obvious in the many studies that clays are dioctahedral and aluminous, it should not show vermiculite increasing towards the surface of be forgotten that trioctahedral micaceous soil clays soil profiles with a concomitant decrease in the do exist. Thus, Fordham (1990a) found that the intensity of the 10 A˚ mica peak. The most detailed weathering of in Australia, in conditions that studies have focused on the biotite?vermiculite led to the formation of lateritic soils, yielded transformation. trioctahedral micaceous clays in the initial stages The structure of vermiculite itself, following of the weathering sequence. Again, Wilson et al. principally upon the fundamental studies of de la (1997) recorded trioctahedral micaceous clays Calle and co-workers (references in de la Calle & accompanied by kaolinite in some alluvial soils Suquet, 1988), is well understood and the structures from Nigeria. In both of these cases the trioctahe- of both the silicate layers and the interlamellar dral clays appeared to be derived from biotite and space, as well as the various ways in which the this raises the question as to why vermiculitization layers may be stacked according to the saturating of the biotite did not occur, as is commonly found cation and relative humidity, have been elucidated. in soils of humid-temperate climates. It may be The vermiculitization of biotite involves a number relevant that in tropical soils, a direct biotite?kao- of factors that have been explored in laboratory linite transformation can occur without an inter- experiments. For example, the release of interlayer vening vermiculite stage (De Kimpe & Tardy, K to an external solution may take place by a 1968). It should also be noted that oxidized biotites diffusion process requiring that the concentration of retain their K more strongly than unoxidized K in the solution is below a critical level of 11 ppm biotites and are hence more resistant to vermiculi- (Newman, 1969). The vermiculitization reaction tization. A possible explanation for the rather ceases if this critical level is exceeded. Numerous unexpected occurrence of trioctahedral micaceous studies have shown that vermiculitization of biotite clays in tropical soils is, therefore, that the also involves the oxidation of structural Fe (Farmer precursor biotites were oxidized and weathered et al., 1971; Vicente-Hernandez et al., 1983) and directly to kaolinite. Such kaolinite formation does that in order to maintain overall electrical neutrality not take place homogeneously throughout the of the structure, this process results in the expulsion biotite crystals, but tends to affect outer surfaces of Fe from octahedral sites. It is not clear whether and planes exposed by exfoliation oxidation of biotite can occur in soils without prior (Ojanuga, 1973) and in these circumstances it is loss of K. Laboratory experiments indicate that loss 12 M. J. Wilson of K is necessary, but the natural occurrence of products, but this proposal does not seem to have oxybiotites as well as evidence that a direct biotite been widely accepted. Fordham (1990b) suggested, to kaolinite weathering step is possible, as indicated on the basis of analytical electron microscopy, that in the above discussion on trioctahedral clay micas, in a granitic weathering profile in Australia, suggest that these experiments may not necessarily trioctahedral vermiculite converted to dioctahedral apply to all soil conditions. vermiculite. It seems unlikely, however, that such a Whether or not previous vermiculitization of mechanism could explain the widespread occur- biotite is essential for oxidation of Fe to occur, the rence of dioctahedral vermiculitic minerals in soils. loss of Fe effectively leads to a more dioctahedral Recent work by Aouidjit et al. (1996), however, structure which in turn leads to a re-orientation of using high resolution transmission electron micro- octahedral hydroxyl from a direction normal to the scopy (HRTEM) presented lattice images showing layer structure in unaltered biotite to a direction that that vermiculitization of muscovitic-type mica did is more inclined to the layer structure in oxidized occur in very fine particles engendered by biotite (Juo & White, 1969). This effectively places microdivision. In particular, they showed micro- the interlayer K into a less negative environment so graphs showing particles with a muscovitic core and that it is consequently held more tightly within the a vermiculitic rim consisting of ~15 layers as well structure (Norrish, 1973). The changes involving as small vermiculitic particles derived from oxidation and loss of Fe have been suggested as consisting of 2 to 5 layers in thickness. explanations for the loss of layer charge in the This work suggests that very fine grained mica biotite to vermiculite transformation, but other particles could be more susceptible to vermiculitiza- mechanisms may also be involved such as loss of tion in the soil by reason of their higher specific hydroxyl ions. surface and enhanced reactivity. Despite the weight of research effort on trioctahedral vermiculites, it is a fact that dioctahe- Smectite minerals dral vermiculite is much more common in soil clays. This mineral, first discovered in British soils The factors that strongly influence the origin and (Brown, 1953) occurs particularly in severely formation of smectites in soils, as reviewed by weathered soils such as Ultisols, Alfisols or even Borchardt (1989), include low-lying topography, Oxisols and is also found commonly in the upper poor drainage and base-rich parent material, leading horizons of Spodosols (Douglas, 1989). Frequently, to favourable chemical conditions characterized by dioctahedral vermiculite is interlayered with non- high pH, high silica activity and an abundance of exchangeable hydroxy Al. The reason dioctahedral basic cations. These conditions are met in many vermiculitic minerals have not been so extensively soils under temperate or cold climates or even in studied is that they are found largely in clay tropical climates where leaching is limited for fractions, where they are difficult to purify. A various reasons, including low precipitation, a muscovite to dioctahedral vermiculite transforma- horizon in the soil profile that impedes the tion analogous to the biotite to trioctahedral passage of water or a naturally high water table vermiculite transformation in the non-clay fractions (Allen & Hajek, 1989). of soils has rarely, if ever, been observed, and Soil smectites are overwhelmingly dioctahedral, mucovitic mica cannot be readily vermiculitized in with only montmorillonite and beidellite being of the laboratory without resort to extreme chemical any real importance. However, from a review of techniques. Thus, there is no ready source of published chemical analyses, Wilson (1987) material with which to study experimentally such concluded that soil smectites were somewhat more problems as the mechanisms involved in the loss of Fe-rich than montmorillonite and beidellite sensu layer charge in dioctahedral vermiculite as there is stricto and tended to show intermediate charge for their trioctahedral counterparts. characteristics with many falling into the category Although the formation of dioctahedral vermicu- of ferruginous beidellites. These observations imply lite from a precursor dioctahedral mica seems that soil smectites could be somewhat different to probable in the majority of cases, it is of interest those associated with bentonite and other geological that other pathways have been suggested. Thus, in deposits, which is perhaps not surprising consid- an early study, Barshad & Kishk (1969) presented ering the range of chemical conditions that may evidence to support formation as precipitation occur in soils as well as the three distinct pathways Origin and formation of clay minerals in soils 13

(inheritance, neoformation and transformation) that silicates are unlikely to be of importance in may be involved in soil smectite formation. Also, in Aridisols. the soil environment there is often no shortage of Smectites derived from inherited materials are Fe, and pedogenic conditions often promote its common in Inceptisols, e.g. in the Fluventic solubilization and mobilization. Haplaquepts of the alluvial Indo-Gangetic plain Wilson (1993) reviewed the literature in order to (Islam & Lotse, 1986), in the Aquepts of the assess the relative importance of the three pathways Bangladesh Holocene floodplain (Brammer & for smectite formation in the Orders of Soil Brinkmann, 1977), in Typic Humaquept on glacial Taxonomy and some of the main points are as marine drift in north west Washington (Pevear et follows. al., 1984) and in Japanese paddy soils derived from In Entisols, smectites are often inherited. Thus, marine alluvium (Egashira & Ohtsubo, 1983). inherited trioctahedral smectite is found in the Neoformational smectites are also common in proto-soils forming from the 1980 Mount Saint Inceptisols, a good example being that described Helens pyroclastic flow (Pevear et al., 1982; La from a Haplaquept in Nigeria by Bui & Wilding Manna & Ugolini, 1987) and inherited smectite (1988) where in the prevailing aquic environment, may occur in Entisols under an extreme range of the silica translocated in the ferruginous soil has climatic conditions as, for example, in the alluvial promoted the formation of an Fe-rich montmorillon- soils of the Blue Nile Clay Plains of Sudan (Wilson ite. Previous work by French scientists (Trauth et & Mitchell, 1979) and in Orthents and Psamments al., 1967; Paquet 1967; Tardy et al., 1974; Pedro et on altered basalt flows in the Faeroe Islands al., 1978) in this area of Africa also suggested a (Rutherford & Debenham, 1981). Neoformational widespread neoformational origin for soil smectite. smectite in Entisols may encompass a similar There seems to be little evidence in the literature climatic range occurring in high pH footslope for smectite to originate in Inceptisols by soils in southern India (Murali et al., 1978), in transformation processes. waterlogged horizons in typic Cryaquents in the Vertisols are dominated by smectite originating Bolivian Andes (Wilke & Zech, 1987) and in through inheritance or neoformation. Vertisols may Torrifluvents formed in wadi deposits under a xeric have formed on alluvial plains where inherited moisture regime in Saudi Arabia (Mashhady et al., smectite is of detrital origin such as in Sudan 1980). Only one example was found where smectite (Wilson & Mitchell, 1979), Turkey (O¨ zkan & Ross, in an Entisol was shown to form by transformation 1979; Guˆzel & Wilson, 1981) and northern Uruguay of a mica mineral, in this case in an (Rossignol, 1983). In these instances, the smectites altered basalt (Reid et al., 1988). This occurrence is are all Fe-rich with a high tetrahedral charge. Similar likely to be exceptional. smectites are found in Vertisols developed upon Smectite may be the dominant clay mineral in basic igneous rocks, such as found in Kenya (Kantor Aridisols in the United States, and Dregne (1976) & Schwertmann, 1974), Israel (Singer, 1971) and suggested that generally, environmental conditions Jordan (Shadfan, 1983) where the clay mineral has a are conducive to smectite formation. Smectite in neoformational origin. Smectites originating by arid soils has been noted in Iraq (Al Ravi et al., transformation of micas have not been conclusively 1969), Iran (Abtahi 1977) and Saudi Arabia (Aba- demonstrated to occur in Vertisols although Badroui Huseyn et al., 1980). Convincing evidence for a & Bloom (1990) and Graham & Southard (1983) neoformational origin for smectite in saprolites have suggested the possibility. However, it would be weathering under arid conditions of the Negev surprising if a mica-to-smectite conversion is able to desert of southern Israel was presented by Singer occur in the poorly drained conditions characteristic (1984). With regard to the smectitic argillic of Vertisols. horizons in the Argid sub-order in the USA, most Smectites are common in Mollisols, particularly evidence suggests that these are Pleistocene relics in the United States. An early study by Glenn et al. related to a more moist climate and to eluvial (1960) concluded that the smectite formed deposition in the soil (Nettleton & Peterson, 1983). following weathering of micas and chlorite and in The smectite would, therefore, be regarded as being a later review, Jackson (1965) concluded that of inherited origin when viewed from the standpoint smectites in many Mid-West soils had been of current pedogenic conditions (Buol, 1965). formed by K depletion of di- and tri-octahedral Smectites originating by transformation of layer micas or by the silication of polymerized sheets of 14 M. J. Wilson hydroxy-Al. In other words, the smectites are moisture regime. Thus, Kantor & Schwertmann largely pedogenic, forming by a combination of (1974) found that smectite increased with depth in transformation and neoformation. In a more recent some Kenyan Humults where it apparently formed study, Laird et al. (1988) investigated smectites in as an initial weathering product of basic igneous an Argialboll-Argiaquoll toposequence in Iowa to rocks only to decompose in the surface horizons. determine whether drainage differences influenced Smectitic Ultisols have been more extensively the chemistry and layer charge of the clay minerals. studied in the USA where they may be developed It was concluded that there was no drainage effect upon smectite-rich , for example in the and that the smectite was likely to be a weathering Alabama Coastal Plains (Karathanasis et al., 1986) product of inherited mica. It should be noted, and in Texas (Carson & Dixon, 1972). The above however, that a comprehensive review of Canadian examples illustrate that Ultisol smectites are in the soils by Kodama (1979) reported that there had main inherited from suitable parent materials or been only minimal transformation of the layer may form, often ephemerally, through neoformation silicate clays in the Mollisols of the Prairie processes at the base of the profile. An origin Provinces, with slight increases in expansible through transformation is unlikely, although this has clays towards the surface in only a few instances. been suggested where Ultisols have formed on In Alfisols, the smectite of the argillic layer is glauconitic sediments in the Mississippi Coastal surmised to have formed by different mechanisms. Plain (Nash et al., 1988). Where the parent material contains illite, it has been In view of the intensively weathered nature of proposed that smectite forms by a transformation Oxisols, it would be expected that smectites would process, with vermiculite as an intermediate phase be unlikely to occur. Certainly, this appears to be in some Ohio Hapludalfs (Smeck et al., 1968; true for the diagnostic oxic horizon or for the Wilding et al., 1971). Generally, however, it has plinthite found at depth. However, when lateritic been proposed that smectite increases in the more profiles are studied in detail, smectite is sometimes poorly drained Aqualfs located in low-lying areas found in the saprolite as an ephemeral phase (Nahon (Allen & Fanning, 1983), presumably being of & Colin, 1982; Singh & Gilkes, 1993) or even in the neoformational origin. Convincing evidence for the so-called pallid zone (Singh & Gilkes, 1991). synthesis of smectite in a Lithic Hapludalf derived Considering the relative importance of the from serpentinite, which emphasized the role of various modes of origin for smectite in Soil poor drainage in the soil, was presented by Istok & Taxonomy, the author’s conclusions are summar- Harward (1982). ized in Table 3. Inheritance from smectitic parent The albic horizon in Spodosols is the site of material is judged to be probably of greatest intense weathering and numerous studies have importance where smectites are found in Entisols demonstrated that it is dominantly smectitic and and Ultisols and of major importance in Aridisols, related to the weathering of mica and sometimes Inceptisols and Vertisols. A neoformation origin is chlorite. Ross & Mortland (1966) first showed that of major importance in Inceptisols, Alfisols, such smectite was beidellitic, a finding since Aridisols and Vertisols and of moderate importance confirmed by others (Churchman, 1980; McDaniel in Entisols and Mollisols. Smectite formed by et al., 1995). The aluminous nature of these clays, transformation is thought to be of major importance as well as the detailed study of Kodama & Brydon only in Spodosols and perhaps also in Mollisols and (1966) which showed that the smectite mineral in Alfisols although the evidence does not appear to albic horizons is a three-component random be entirely conclusive. As a general comment, it interstratification of mica, vermiculite and smectite, seems to the author that too much importance may represents convincing evidence that the smectite have been placed upon transformation mechanisms originates through transformation of the mica. A as a means of explaining the origin of soil smectites chlorite to smectite transformation is also possible and not enough on neoformation and inheritance. as Spodosols with smectitic E horizons are some- times found on parent materials which contain Chlorites and interlayered minerals chlorite but no mica such as the tephras of the Pacific north west USA (Ugolini et al., 1991). As indicated by Barnhisel & Bertsch (1989) there Smectites are usually found as a transitory phase is no doubt that the occurrence in soils of primary in Ultisols particularly where there is an aquic chlorite of ferromagnesian, trioctahedral nature, is Origin and formation of clay minerals in soils 15

TABLE 3. Origin of soil smectites in soil orders as assessed from the literature.

Soil orders Inheritance Neoformation Transformation

Entisols +++ ++ + Aridisols +++ +++ À Inceptisols +++ +++ ++ Vertisols +++ +++ + Mollisols + ++ ++ Alfisols + +++ ++ Spodosols À + +++ Ultisols +++ + À Oxisols ÀÀÀ

+++ Major importance ++ Moderate importance + Minor importance À No importance due largely to inheritance. Primary chlorites are by Vicente et al. (1977) to result in the formation of generally considered to be easily weatherable and Al hydroxy interlayered vermiculite, whereas the Ross & Kodama (1976) provided convincing effect of highly complexing organic acids often experimental evidence to show that chlorite could resulted in complete breakdown of the mineral be changed into a regularly interstratified chlorite- structure. Many experiments have shown that water- vermiculite through chemical oxidation and selec- soluble polynuclear Al complexes are strongly tive extraction of the interlayer hydroxide sheet. adsorbed by expansible minerals to yield hydroxy- Field occurrences of the vermiculitization of interlayered clays (e.g. de Villiers & Jackson, chlorite have been documented by Proust (1982), 1967), and that the original exchangeable cations Proust et al. (1986) and by Buurman et al. (1988). are quantitatively displaced (Brown & Newman, Alternatively, chlorite may be completely decom- 1973). According to Vicente et al. (1977) Al posed without the formation of any crystalline hydroxy-interlayered clays in soils mainly relate product, as found by Bain (1977) in podzolic soils directly to the decomposition of the layer silicate in Scotland. mineral itself. They further concluded that inter- The origin of so-called pedogenic chlorite and layered minerals are the most common clay interlayered or inter-gradient minerals appears to be minerals in the soils of Northern Europe and due entirely to transformation reactions involving North America and that they may result not only the introduction of non-exchangeable hydroxy-Al from natural soil processes but may also be polymers into the interlamellar space of pre-existing promoted by the application of fertilizers and the smectite, vermiculite or interstratified expansible deposition of acidity from the atmosphere. minerals. The source of the interlayer Al could be Extensive consideration has been given to these direct involving acid attack of both tetrahedral and minerals in the latter context, particularly because octahedral sheets of the layer , or the stability of the interlayers is pH dependent, indirect from the soil solution following weathering being favoured in the pH range extending of aluminous minerals, like feldspars, or from approximately from 4.0 to 5.8 (Barnhisel & decomposition of organic matter containing Bertsch, 1989). Hydroxy interlayered clays can adsorbed Al. Both pathways, as well as the thus serve as sources and sinks of mobile Al conditions under which they operate, have been depending upon soil pH conditions. Other soil confirmed by laboratory experiments. For example, factors that promote their formation include low it has long been known that acidified smectite organic matter content, oxidizing conditions and spontaneously converts to an Al-saturated form frequent wetting and drying cycles. These condi- whereby H+ is replaced by structural Al. Again, the tions are met in a wide range of soils, and decomposition of trioctahedral micas by mineral interlayered minerals are thus commonly encoun- acids and non-complexing organic acids was shown tered. They may be most typical of highly 16 M. J. Wilson weathered soils and they certainly occur in Ultisols of soil types. However, because they represent (Karathanasis, 1988) and even in Oxisols. intermediate transformation products, they would However, some emphasis has recently been perhaps not be expected to occur widely in old placed on the occurrence of hydroxy interlayered highly weathered soils. clays in the podzolic soils of catchments on slowly Wilson & Nadeau (1985) reviewed the types of weathering parent rocks and where surface waters interstratified minerals that are characteristic of are considered to be at risk from acidification. For weathering environments. Hydrobiotite, consisting example, April et al. (1986) described hydroxy Al of large crystals of regularly alternating biotite and interlayered vermiculite, primarily in B horizons of vermiculite layers and derived largely from the soils from 26 Adirondack catchments. Olson (1988) weathering of biotite, represents a classic case of an suggested that such minerals play a key role in interstratified mineral that was observed early on. It neutralizing potential increases in acid deposition in has been found many times in soils (e.g. Coleman catchments in montane areas of Maryland and et al., 1963; Wilson, 1970; Kapoor, 1972), appears Virginia. In Scotland, Bain et al. (1990) presented to form relatively rapidly (Wilson & Nadeau, evidence to show the occurrence of these minerals 1985), and can be synthesized in the laboratory in the podzolic soils of three upland catchments and using dilute solutions to exchange interlayer K also demonstrated the pH dependency of the (Rausel Colom et al., 1965) or by oxidation of interlayers. The extent of the interlayering declined structural Fe during cation exchange (Farmer & with falling pH and complete removal was detected Wilson, 1970). Norrish (1973) proposed a convin- below pH 4.3 in the E horizons. cing mechanism for the formation of hydrobiotite Whether acid precipitation has a major role in involving sympathetic re-orientation of the hydro- bringing about such changes is rather difficult to xyls on either side of the octahedral sheet. Removal determine, although from laboratory experiments it of K from one side of the silicate layer increases is known that changes in hydroxy interlayering of the angle of the (OH) bond direction on that side of expansible phases can occur very rapidly. In most the octahedral sheet, but causes a decrease in the soil studies there is no historical time scale to angle of the (OH) band on the other side. This assess the rate of change and to judge whether results in the interlayer K on this side being placed changes coincide with marked periods of acid in a more negatively charged and hence more stable deposition. environment and alternating interlayer regions are Hydroxy Al interlayers in expansible clay built up within single crystals where K is either minerals may become silicated as a result of depleted or held very strongly. Regularly inter- interactions either outside or within the interla- stratified dioctahedral mica-vermiculite is also mellar space (Lou & Huang, 1993) leading to the found in soil clay fractions particularly in the formation of an intergradient vermiculite-kaolin upper horizons of podzolic soils (Churchman, mineral (Wada & Kakuto, 1983) or to kaolinite 1978,1980; Wilson et al., 1984) and presumably directly (Karathanasis & Hajek, 1983) as will be formed from dioctahedral mica by a mechanism discussed below. Interlayered minerals are mainly similar to that of hydrobiotite formation. However, dioctahedral, representing the nature of the the dioctahedral mineral does not seem to occur in precursor expansible minerals, but Al interlayered non-clay fractions. Further weathering of vermicu- trioctahedral vermiculites may also occur during litized mica results in a lower negative charge on biotite weathering (Kato, 1965; Wilson, 1965,1966). the silicate sheet leading to the formation of smectitic products of both a trioctahedral (Ismail, 1969; MacEwan, 1954) and dioctahedral nature Interstratified minerals (Churchman, 1980). In the latter instance a These minerals primarily represent intermediate regularly interstratified mineral may form with the transformation products, mainly involving mica, XRD characteristics of rectorite. chlorite and an expansible phase, either smectite or Regularly interstratified chloritic material may vermiculite (Sawhney, 1989), although there are occur in soils as a consequence of the vermiculitiza- many examples of interstratified minerals in soils tion of chlorite, although the mechanism involved is originating by inheritance. Both regularly and not clear. A regularly interstratified chlorite- randomly interstratified clays either of dioctahedral vermiculite yielding a high spacing of 28 A˚ with or trioctahedral nature, may occur in a wide variety many lower orders was described by Johnson Origin and formation of clay minerals in soils 17

(1964) from the C horizon of a soil developed on & Grubb, 1991). The transformation of smectite to metamorphosed basalt in Pennsylvania and a similar interstratified halloysite-smectite was shown in mineral was reported by Herbillon & Makumbi weathered acid clay in Japan to involve the (1975) in a recent tropical soil derived from crystallization of excess silica in a separate phase chlorite- in Zaı¨re. In the latter instance, at (Watanabe et al., 1992). least, it was evident that the chlorite from which the Although Wilson (1987) concluded that the interstratified product developed was already partly concept of fundamental particles was probably not vermiculitized (Makumbi & Herbillon, 1972) but of general applicability in the interpretation of the experiments of Ross & Kodama (1976) interstratified smectitic clays in soils, Robert et al. demonstrate that some true chlorites will break (1991) showed by HRTEM that these particles did down to a regularly interstratified product. exist in soils and, from these observations, Aouidjit The formation of regularly interstratified minerals et al. (1996) distinguished two types of inter- and thence of vermiculite or smectite must involve stratification described as ‘structural’ and ‘textural’. transition through randomly interstratified phases The former involves relatively large crystals of but it is of interest that these phases appear to be micrometric lateral extent, variable thickness and relatively transitory and do not seem to persist in three-dimensional structure. Textural interstratifica- soil profiles. Comparison with diagenetic sequences tion consists of associations of elementary particles is striking where great thicknesses of argillaceous with a random two-dimensional structure and is of sediment may be characterized by randomly variable thickness and lateral extent. interstratified illite-smectite which converts slowly with depth and increasing temperature and pressure Kaolin minerals to more ordered forms including rectorite-like structures (R1) and more long-range ordered Dixon (1989) reviewed the origin and formation structures (R2 and R3). In the author’s view, there of kaolin minerals and showed that precipitation is little evidence that the long-range ordered illite- from solution of kaolinite required acid conditions smectites nor indeed the precursor randomly with moderate silica activity and small amounts of interstratified illite-smectite clay form as stable base cations. Kaolinite was synthesized from phases in soils. This is perhaps not surprising in hydroxy Al interlayered montmorillonite by view of the increasing PÀT conditions required for Poncelet & Brindley (1967) but only under this transformation in diagenesis. One of the few hydrothermal conditions. Neoformation and trans- papers to show that interstratified illite-smectite will formation mechanisms are, therefore, feasible. form at earth surface conditions is that of Bergkraut Kaolinite is generally found as a minor constituent et al. (1994) in a study of weathering in basic in young soils, such as those derived from glacial pyroclastics. However, the illite-smectite differs material, and by implication is considered to require from that found in diagenetic sequences in that it >10,000 years to form. It may be an abundant is Fe-rich, recalling the illitic clays formed at earth constituent of soils developed upon old geomorphic surface conditions. surfaces. Interstratified kaolin-smectite is now becoming Compared with the of geological more widely reported in soils. It appears to be deposits, soil kaolinites are often of a smaller particularly common in soils derived from basic particle size, tend to be highly disordered and may igneous rocks and, for example, was identified by contain Fe as an isomorphous substituent. They Norrish & Pickering (1983) as the only clay mineral may also be interstratified with smectites. The in ~40 Australian soils developed upon such parent pioneering work of Angel et al. (1974), Jones et al. material. The occurrence of kaolinite-smectite in a (1974) and Meads & Malden (1975) using EPR red-black toposequence in Burundi (Herbillon et al., spectroscopy showed that Fe3+ could be substituted 1981) suggests a genetic link between the black for Al in the octahedral sheet of kaolinite and that smectitic soils in the poorly drained footslopes and there were different sites of substitution. For the the red kaolinitic soils on the freely-drained hill kaolinite of tropical soils Herbillon et al. (1976) tops. More recent work shows that the kaolin demonstrated a relationship between Fe content, mineral in this kind of interstratification may be crystallinity index and EPR spectral features, halloysite (Delvaux et al., 1990) and that the particularly the intensity and asymmetry of the expansible mineral can be vermiculitic (Bu¨hmann geff~4 band. Mendelovici et al. (1979) and 18 M. J. Wilson

Cantinolle et al. (1984) showed convincingly by IR there is evidence that a transformation mechanism spectroscopy that structural Fe occurs in kaolinite in is feasible. Thus, Wada & Kakuto (1983) presented lateritic and bauxitic soils. It seems, therefore, that evidence to show that in Korean Ultisols, kaolinite soil kaolinites (like soil smectites), may be different had formed via an intergradient vermiculite-kaolin from the kaolinites of geological deposits, presum- mineral involving tetrahedral inversion and attach- ably because of the large amounts of Fe in the ment to the hydroxy Al interlayer sheet. A similar pedogenic environment combined with conditions mechanism for the formation of kaolinite from that promote the mobilization of this element, at hydroxy interlayered smectite was proposed by least on a local scale. The characterization of Karathanasis & Hajek (1983). kaolinite by EPR was developed further by Mu¨ller & Calas (1993) who showed that the technique FUTURE PERSPECTIVES could be used to distinguish different environments of formation according to total Fe3+ content, This overview has highlighted the necessity for distribution among different sites (Fe(I)and further research to address many specific questions Fe(II)), concentration and types of paramagnetic and gaps in our knowledge as to the origin and defect centres (PDC) and presence or absence of formation of clay minerals in soils, as well as sorbed species. Soil kaolinites yielded a high further characterization of the detailed nature of the intensity Fe3+ EPR signal, showed an interdepen- minerals themselves. dence between Fe(I) concentrations and crystalline With regard to soil micas, it is clear that further order, low intensity signals from PDCs and lack of detailed studies are needed on their crystallochem- resonance due to sorbed species. istry and on the changes that occur with decreasing Halloysite occurs primarily in youthful volcanic- particle size, particularly the possible conversion derived soils, but it also forms from primary from the 2M1 to the 1Md polytype and the role of minerals in tropical soils or pre-glacially weathered microdivision in bringing about fundamental materials. Most commonly the mineral is of a particles, or at least particles that are more tubular or spheroidal morphology. Platy halloysite susceptible to vermiculitization. At this stage the is rare. Although similar to kaolinite in its layer relative roles of increased basal surface as opposed structure, it is important to realize that halloysite is to edge surface is unclear. The question of fundamentally different in that it has a 2-layer formation of pedogenic micas is one that remains structure whereas kaolinite has a 1-layer structure to be resolved. Even the reality of the process (Bailey, 1989). This means that a halloysite to seems uncertain, a point that could be clarified by kaolinite conversion cannot take place by a simple K/Ar dating and other studies of carefully structural transformation but would require a fractionated samples. recrystallization process, as was found experimen- For reasons that have been given above, most of tally by La Iglesia & Gala´n (1975). The reasons for the information that we have on vermiculite refers the formation of spheroidal halloysite as opposed to to trioctahedral vermiculite, despite the fact that tubular halloysite are not clear, although it may be dioctahedral vermiculite is much more common in that high levels of silica supersaturation are soils. Therefore, we need to know more about the required (Tomura et al., 1985) and the appropriate mechanisms of formation of dioctahedral vermicu- physical conditions provided by pumice grains. lites, particularly the changes in chemistry involved Tubular halloysite may eventually form from the in the reduction of layer charge during evolution spheroidal form (Sudo & Yotsumoto, 1977). It is from the parent micas. In this instance, oxidation of worth noting too that, like kaolinite, halloysite may structural Fe is unlikely to play as significant a role contain structural Fe (Wada & Mizota, 1982), as it does in trioctahedral vermiculites. There is particularly where it has formed from the weath- very little information on the stacking sequence of ering of ferruginous material (Dao Cho & Mermut, the silicate layers and the organization of the 1992). There is evidence gained by X-ray photo- interlamellar space for dioctahedral vermiculites, electron spectroscopy that spheroidal halloysite may in comparison with that available for their have a greater Fe content than the tubular form trioctahedral counterparts. (Bailey, 1989). Although there is good evidence that soil Although most studies would seem to indicate smectites may form through both transformation that kaolinite in soils has formed by neoformation, and neoformation pathways, more information is Origin and formation of clay minerals in soils 19 needed with regard to possible differences in For the kaolin minerals, it seems that genesis chemistry and structure that these two modes of through a transformation mechanism from hydroxy- origin may engender. For example, transformation interlayered minerals involving tetrahedral inversion smectites should, in principle, be beidellitic, as they of silicate layers requires further documentation and should inherit Al-substituted tetrahedral sheets from assessment as to its general applicability in soils. their parent micas, and they might also be expected Also, there is no question that a better under- to show semi-random stacking (again an inherited standing is required of the conditions of formation feature from their precursor vermiculites), as of halloysite as opposed to kaolinite, in soils, and in opposed to turbostratic stacking. The mechanism particular the formation of spheroidal halloysite vs. involved in the reduction of layer charge, and the tubular halloysite. In these endeavours, it seems distinction from the vermiculite from which the certain that the use of advanced spectroscopic transformation smectites are derived is another area techniques such as EPR will play an ever increasing for further research. With regard to neoformation role. smectites, confirmation is required by direct With regard to the global overviews of clay methods that many can indeed be described as mineral formation in soils described at the ferruginous beidellites, as is suggested by chemical beginning of this paper, it still seems that they analyses. However, such analyses can be misleading can be regarded as generally valid and that they because non-structural Fe oxides may be resistant to provide a useful frame of reference. It may be, deferration techniques. There may perhaps be scope however, that uncertainty still remains about the to use EPR techniques for the assessment of relative importance of transformation and neofor- environmental conditions for smectite formation, mation in both general and specific instances. in the same way as they have been used in kaolinite Figure 2 summarizes the author’s views on this studies. In any event, it would seem that detailed matter in terms of primary and secondary pathways comparison between transformation and neoforma- for clay mineral formation in soils in a global tion soil smectites would be worthwhile. context. It can be seen that soil smectites and In the author’s view, there is ample scope for kaolinites are regarded primarily as the products of further research into the interstratified clay minerals neoformation. For smectite, an important, but that are truly characteristic of the pedogenic secondary pathway, applicable mainly to environment. It seems that the present picture is Spodosols and perhaps to Mollisols and Alfisols, somewhat confused by the implied formation in is through the weathering of inherited mica (and to soils of the same types of interstratified illite- a lesser extent chlorite) and may or may not involve smectites that are so common in the diagenetic the intervention of hydroxy-Al interlayered environment. This is unreasonable, bearing in mind minerals. Kaolinite too, although primarily a soil that the latter are characterized by increasing mineral arising from neoformation, also forms conditions of temperature and pressure and that through other pathways of transformation, namely there is a dearth of evidence to show that direct weathering of mica, conversion of hydroxy- diagenetic-type illite-smectites really do form in Al interlayered vermiculite or smectite and weath- the soil. It is true that K saturation combined with ering of smectite through an intermediate kaolinite- wetting and drying cycles can convert smectite to smectite interstratified phase. Halloysite is regarded an illite-like product in the laboratory, but this primarily as a product of neoformation, although it product is not illite sensu stricto and K saturation is seems that it may also form through an inter- unlikely to occur in the soil except under arid stratified halloysite-smectite as a result of smectite conditions. Further research is required to address weathering. this issue. Other specific problems that might be Finally, it is worth emphasizing that the question addressed include the mechanisms for the formation of the origin and formation of clay minerals in soils of regularly interstratified dioctahedral mica-vermi- is not an exercise of mere academic interest but culite and for interstratified kaolinite-smectite and may relate directly to soil properties and behaviour halloysite-smectite in soils. Further information is in numerous ways. A better understanding of the required too on the possible role of microdivision detailed nature of clay minerals in soil and the way and the creation of fundamental particles in in which they relate to overall soil chemical and promoting interstratification as suggested by physical properties, combined with a knowledge of Robert and his colleagues. their mode of formation, provides a sound basis for 20 M. J. Wilson

FIG. 2. Pathways for the formation of clay minerals in soils. Note that the major pathways for smectite and kaolinite are considered to be through neoformation as represented by thick arrows. Transformation pathways to smectite occur via interstratified and interlayered phases. A direct conversion of mica to kaolinite can also occur but this may not be regarded as a true transformation because of the structural differences between the minerals, hence the dotted line. being able to predict, at least in general terms, soil REFERENCES behaviour in the context of both agricultural and Aba-Huseyn M.M., Dixon J.B. & Lee S.Y. (1980) environmental problems. Mineralogy of Saudi Arabian soils: south western Region. Soil Sci. Soc. Am. J. 44, 643À649. Abtahi A. (1977) Effect of saline and alkaline ground ACKNOWLEDGMENTS water on soil genesis in semi arid Southern Iran. Soil I am indebted to the Clay Minerals Group Committee Sci. Soc. Am. J. 41, 583À588. for the invitation to prepare this review paper. This Allen B.L. & Fanning D.S. (1983) Comparison and Soil work was supported by the Scottish Office Agriculture Genesis. Pp. 141À192 in: Pedogenesis and Soil Environment and Fisheries Department. Taxonomy. I. Concepts and Interactions (L.P. Origin and formation of clay minerals in soils 21

Wilding, N.E. Smeck & G.F. Hall, editors). Elsevier, 24, 339À354. Amsterdam. Bu¨hmann C. & Grubb P.L.C. (1991) A kaolinite Al Ravi A.H., Jackson M.L. & Hole F.D. (1969) smectite interstratification sequence from a red-black Mineralogy of some arid and semi-arid land soils toposequence. Clay Miner. 26, 343À358. of Iraq. Soil Sci. 107, 480À486. Bui E. & Wilding L.P. (1988) Pedogenesis and Allen B.L. & Hajek B.F. (1989) Mineral Occurrence in mineralogy of a Haplaquept in Niger (West Soil Environments. Pp. 199À278 in: Minerals in Soil Africa). Geoderma, 43,49À64. Environments (J.B. Dixon & S.B. Weed, editors). Buol S.W. (1965) Present soil-forming factors and Soil Sci. Soc. America, Madison, Wisconsin, USA. processes in arid and semi arid regions. Soil Sci. Angel B.R., Jones J.P.E. & Hall P.L. (1974) Electron 99,45À49. spin resonance studies of doped synthetic kaolinite. Buurman P., Meyer E.L. & Van Wijk J.H. (1988) I. Clay Miner. 10, 247À255. Weathering of chlorite and vermiculite in ultramafic Aouidjit H., Elsass F., Righi D. & Robert M. (1996) rocks of Cabo Ortegal, northwest Spain. Clays Clay Mica weathering in acidic soils by analytical electron Miner. 36, 263À269. microscopy. Clay Miner. 31, 319À332. Calle de la C. & Suquet H. (1988) Vermiculite. Pp. April R.H., Hluchy M.M. & Newton R.M. (1986) The 455À496 in: Hydrous Phyllosilicates (exclusive of nature of vermiculite in Adirondack soils and till. micas) (S.W. Bailey, editor). Vol. 19. Reviews in Clays Clay Miner. 34, 549À556. Mineralogy. Mineralogical Soc. America, Badraoui M. & Bloom P.R. (1990) rich high charge Washington. beidellite in Vertisols and Mollisols of the High Cantinolle P., Didier P., Meunier J.D., Parron C., Chaouia Region of Morocco. Soil Sci. Soc. Am. J. 54, Guendon J.L., Bocquier G. & Nahon D. (1984) 267À274. Kaolinites ferriferes et oxy-hydroxydes de fer et Bailey S.W. (1989) Halloysite — a critical assessment. d’alumine dans les bauxites des Cantonnettes (S E de Proc. 9th Int. Clay Conf. Strasbourg,89À98. la France). Clay Miner. 19, 125À135. Bain D.C. (1977) The weathering of ferruginous chlorite Carson C.D. & Dixon J.B. (1972) selectivity in a podzol from Argyllshire, Scotland. Geoderma, in certain montmorillonitic soil clays. Soil Sci. Soc. 17, 193À208. Am. Proc. 36, 838À843. Bain D.C., Mellor A. & Wilson M.J. (1990) Nature and Churchman G.J. (1978) Studies on a climax sequence in origin of an aluminous vermiculitic weathering soils in tussock grasslands. Mineralogy N. Z. J. Sci. product in acid soils from upland catchments in 21, 467À480. Scotland. Clay Miner. 25, 467À475. Churchman G.J. (1980) Clay minerals formed from Barnhisel R.I. & Bertsch P.M. (1989) Chlorites and micas and chlorites in some New Zealand soils. Clay Hydroxy Interlayered Vermiculite and Smectite. Pp. Miner. 15,59À76. 729À788 in: Minerals in Soil Environments (J.B. Coleman N.T., LeRoux F.H. & Cady J.G. (1963) Dixon & S.B. Weed, editors). Soil Sci. Soc. Amer., Biotite-hydrobiotite-vermiculite in soils. Nature Madison, Wisconsin. Lond. 198, 409À410. Barshad I. & Kishk F.M. (1969) Chemical composition Dao Cho Lt. & Mermut A.R. (1992) Evidence for of soil vermiculite clays as related to their genesis. halloysite formation from weathering of ferruginous Contrib. Mineral. Pet. 24, 136À155. chlorite. Clays Clay Miner. 40, 608À619. Basham I.R. (1974) Mineralogical changes associated De Kimpe C. & Tardy Y. (1968) Etude de l’alteration with deep weathering of gabbro in . d’une biotite en kaolinite par spectroscopie infra- Clay Miner. 10, 189À202. rouge. Bull. Groupe Franc. Argiles, 19,81À85. Bergkraut V., Singer A. & Stahr K. (1994) Palagonite Delvaux B., Herbillon A.J., Vielvoye L. & Mestdagh reconsidered: paracrystalline illite-smectite from M.M. (1990) Surface properties and clay mineralogy regoliths on basic pyroclastics. Clays Clay Miner. of hydrated halloysite soil clays. II. Evidence for the 42, 582À592. presence of halloysite smectite (H/Sm) mixed layer Borchardt G. (1989) Smectites. Pp. 675À727 in: clays. Clay Miner. 25, 141À160. Minerals in Soil Environments (J.B. Dixon & S.B. Dixon J.B. (1989) Kaolin and Serpentine Group Weed, editors). Soil Sci. Soc. America Madison, Minerals. Pp. 467À525 in: Minerals in Soil Wisconsin. Environments (J.B. Dixon & S.B. Weed, editors). Brammer H. & Brinkman R. (1977) Surface water gley Soil Sci. Soc. America, Madison, Wisconsin. soils in Bangladesh: Environment landforms and soil Dixon J.B. & Weed S.B. (1989) Minerals in Soil morphology. Geoderma, 17,91À109. Environments. Soil Sci. Soc. America, Madison, Brown G. (1953) The dioctahedral analogue of vermi- Wisconsin. culite. Clay Miner. Bull. 2,64À69. Douglas L.A. (1989) Vermiculites. Pp. 635À674 in: Brown G. & Newman A.C.D. (1973) The reactions of Minerals in Soil Environments (J.B. Dixon & S.B. soluble aluminium with montmorillonite. J. Soil Sci. Weed, editors). Soil Sci. Soc. America, Madison, 22 M. J. Wilson

Wisconsin. Istok J.D. & Harward M.E. (1982) Influence of soil Dregne H.E. (1976) Soils of Arid Regions. Elsevier, moisture on smectite formation in soils derived from New York. serpentinite. Soil Sci. Soc. Am. J. 46, 1106À1108. Duchaufour P. (1960) Pre´cis de Pedologie. Masson, Jackson M.L. (1964) Chemical composition of soils. Pp. Paris. 71À141 in: Chemistry of the Soil (F.E. Bear, editor). Dymond J., Biscaye P.E. & Rex R.W. (1974) Eolian Reinhold Publishing Corp., New York. origin of mica in Hawaiian soils. Geol. Soc. Am. Jackson M.L. (1965) Clay transformations in soil Bull. 85,37À40. genesis during the Quaternary. Soil Sci. 99,15À21. Egashira K. & Ohtsubo M. (1983) Swelling and Johnson J.J. (1964) Occurrence of regularly interstrati- mineralogy of smectites in paddy soils derived from fied chlorite-vermiculite as a weathering product of marine alluvium, Japan. Geoderma, 29, 119À127. chlorite in a soil. Am. Miner. 49, 556À572. Fanning D.S., Keramidas V.Z. & El-Desoky M.A. Jones J.P.E., Angel B.R. & Hall P.L. (1974) Electron (1989) Micas. Pp. 551À634 in: Minerals in Soil spin resonance studies of doped synthetic kaolinite. Environments (J.B. Dixon & S.B. Weed, editors). II. Clay Miner. 10, 257À270. Soil Sci. Soc. America, Madison, Wisconsin. Juang T.C. & Uehara G. (1968) Mica genesis in Farmer V.C. & Wilson M.J. (1970) Experimental Hawaiian soils. Soil Sci. Soc. Am. Proc. 32,31À35. conversion of biotite to hydrobiotite. Nature Lond. Juo J.S.R. & White J.L. (1969) Orientation of the dipole 226, 841À842. moments of hydroxyl groups in oxidized and Farmer V.C., Russell J.D., McHardy W.J., Newman unoxidized biotite. Science, 165, 804À805. A.C.D., Alrichs J.L. & Rimsaite J.Y.H. (1971) Kantor W. & Schwertmann U. (1974) Mineralogy and Evidence for loss of protons and octahedral iron genesis of clays in red-black soil toposequences on from oxidized biotites and vermiculites. Mineral. basic igneous rocks in Kenya. J. Soil Sci. 25,63À78. Mag. 38, 121À137. Kapoor B.S. (1972) Weathering of micaceous clays in Fordham A.W. (1990a) Formation of trioctahedral illite some Norwegian podzols. Clay Miner. 9, 383À394. from biotite in a soil profile over granite . Karathanasis A.D. (1988) Compositional and solubility Clays Clay Miner. 38, 187À195. relationships between aluminium-hydroxy inter- Fordham A.W. (1990b) Weathering of biotite into layered soil smectites and vermiculites. Soil Sci. dioctahedral clay minerals. Clay Miner. 25,51À63. Soc. Am. J. 52, 1500À1508. Glenn R.C., Jackson M.L., Hole F.D. & Lee G.B. (1960) Karathanasis A.D. & Hajek B.F. (1984) Evaluation of Chemical weathering of layer silicate clays in loess- aluminium-smectite equilibria in naturally acid soils. derived Tama silt loam of southwestern Wisconsin. Soil Sci. Soc. Am. J. 48, 413À417. Clays Clay Miner. 8,63À83. Karathanasis A.D., Hurt G.W. & Hajek B.F. (1986) Graham R.C. & Southard A.R. (1983) Genesis of a Properties and classification of montmorillonite-rich Vertisol and an associated Mollisol in Northern Hapludults in the Alabama coastal plains. Soil Sci. Utah. Soil Sci. Soc. Am. J. 47, 552À559. 142,76À82. Gu¨zel N. & Wilson M.J. (1981) Clay mineral studies of Kato K. (1965) Mineralogical study of weathering a soil chronosequence in southern Turkey. products of at Shinshiro city (V) Geoderma, 25, 113À129. Trioctahedral aluminium vermiculite as a weathering Herbillon A.J. & Makumbi M.V. (1975) Weathering of product of biotite. Soil Sci. Plant Nutrition, Tokyo, chlorite in a soil derived from a chlorite-schist under 11, 114À122. humid tropical conditions. Geoderma, 13,89À104. Kodama H. (1979) Clay minerals in Canadian soils: Herbillon A.J., Mestdagh M.M., Vielvoye L. & their origin, distribution and alteration. Can. J. Soil Derouane E.G. (1976) Iron in kaolinites with special Sci. 59,37À58. reference to kaolinites from tropical soils. Clay Kodama H. & Brydon J.E. (1966) Interstratified Miner. 11, 201À220. montmorillonite-mica clays from sub-soils of the Herbillon A.J., Frankort R. & Vielvoye L. (1981) An Prairie Provinces, Western Canada. Clays Clay occurrence of interstratified kaolinite-smectite Miner. 13, 151À173. minerals in a red-black soil toposequence. Clay La Iglesia A. & Gala´n E. (1975) Halloysite-kaolinite Miner. 16, 195À201. transformation at room temperature. Proc. Int. Clay Islam A.K.M.E. & Lotse E.G. (1986) Quantitative Conf. Madrid, 173À185. mineralogical analysis of some Bangladesh soils La Manna J.M. & Ugolini F.C. (1987) Trioctahedral with X-ray, exchange and selective dissolution vermiculite in a 1980 pyroclastic flow, Mt St Helens, techniques. Clay Miner. 21,31À42. Washington. Soil Sci. 143, 162À167. Ismail G.T. (1969) Role of ferrous iron oxidation in the Laird D.A. & Nater E.A. (1993) Nature of the illitic alteration of biotite and its effect on the type of clay phase associated with randomly interstratified smec- minerals formed in soils of arid and humid regions. tite-illite in soils. Clays Clay Miner. 41, 280À287. Am. Miner. 54, 1460À1466. Laird D.A., Fenton T.E. & Scott A.D. (1988) Low Origin and formation of clay minerals in soils 23

charge of smectites in an Argialboll-Agriaquoll 165À215 in: Pedogenesis and Soil Taxonomy. II. Sequence. Soil Sci. Soc. Am. J. 52, 4630467. The Soil Orders (L.P. Wilding, N.E. Smeck & G.F. Lou G. & Huang P.M. (1993) Silication of hydroxy Hall, editors). Elsevier, Amsterdam. aluminium interlayers in smectite. Clays Clay Miner. Newman A.C.D. (1969) Cation exchange properties of 41,38À44. micas. I. The relation between mica composition and MacEwan D.M.C. (1954) ‘Cardenite’, a trioctahedral K-exchange in solutions of different pH. J. Soil Sci. montmorillonoid derived from biotite. Clay Miner. 20, 357À373. Bull. 2, 120À126. Niederbudde E.A. (1975) Veranderungen von Mahjoory R.A. (1975) Clay mineralogy, physical and Dreischicht-Tonmineralen durch natives K in holo- chemical properties of some soils in arid regions of za¨nen Mittel Deutschlands und Niederbayerns. Z. Iran. Soil Sci. Soc. Am. Proc. 39, 1157À1164. Pflanzennernaehr. Bodenk. 138, 217À234. Makumbi M.N. & Herbillon A.J. (1972) Niederbudde E.A. & Kussmaul H. (1978) Tonmineral Vermiculitisation e´xperimentale d’une chlorite. Bull eigenschaften und Umwandlungen in Parabraunerde- Groupe Franc. Argiles, 24, 153À164. Profilpaeren unter Acker und Wald in Martı´n-Garcı´a J.M., Delgado G., Sa´ndez-Maron˜o´n M., Su¨ddeutschland. Geoderma, 20, 239À255. Pa´rraga J.F. & Delgado R. (1997) Nature of Norrish K. (1973) Factors in the weathering of mica to dioctahedral micas in Spanish red soils. Clay vermiculite. Proc. Int. Clay Conf., Madrid, Miner. 32, 107À122. 417À432. Masshady A.S., Reda M., Wilson M.J. & Mackenzie Norrish K. & Pickering J.G. (1983) Clay Minerals. Pp. R.C. (1980) Clay and silt mineralogy of some soils 282À308 in: Soils: an Australian Viewpoint: from Qasim, Saudi Arabia. J. Soil Sci. 31, 101À115. Division of Soils, CSIRO. CSIRO, Melbourne/ McDaniel P.A., Falen A.L., Tice K.R., Graham R.C. & Academic Press, London. Fendorf S.E. (1995) Beidellite in E horizons of Ojanuga A.G. (1973) Weathering of biotite in soils of a Northern Idaho spodosols formed in volcanic ash. humid tropical climte. Soil Sci. Soc. Am. Proc. 37, Clays Clay Miner. 43, 525À532. 644À646. Meads R.E. & Malden P.J. (1975) Electron spin Olson C.G. (1988) Clay mineral contribution to the resonance in natural kaolinites containing Fe3+ and weathering mechanisms in two contrasting water- other transition metal ions. Clay Miner. 10, sheds. J. Soil Sci. 39, 457À468. 313À345. O¨ zkan A. & Ross G.J. (1979) Ferruginous beidellites in Mendelovici E., Yariv Sh. & Villalba R. (1979) Iron- Turkish soils. Soil Sci. Soc. Am. J. 43, 1242À1248. bearing kaolinite in Venezuelan laterites. I. Infrared Paquet H. (1967) Les des vertisols: spectroscopy and chemical dissolution evidence. alteration alcaline en milieu tropicale. Bull. Serv. Clay Miner. 14, 323À331. Carte-Geol Als. Lorr. 20, 293À306. Millot G. (1964) Geologie des Argiles. Masson, Paris. Pedro G. (1964) Contribution a l’e´tude expe´rimentale de Muller J.P. & Calas G. (1993) Genetic conditions of l’altere´ration geochimique des roches crystallines. paramagnetic centers in kaolinites. Pp. 261À289 in: The`se Paris, Ann Agron. Kaolin Genesis and Utilization (H. Murray, W. Pedro G. (1982) The conditions of formation of Bundy & C. Harvey, editors). Clay Miner. Soc., secondary constituents. Pp. 63À81 in: Constituents Boulder, Colorado. and Properties of Soils (M. Bonneau & B. Souchier, Murali V., Krishna Murti G.S.R. & Sarma V.A.K. editors) Academic Press, London. (1978) Clay mineral distribution in two topose- Pedro G., Carmouze J.-P. & Velde B. (1978) Peloidal quences of tropical soils of India. Geoderma, 20, formation in recent sediments of Lake 257À277. Chad. Chem. Geol. 23, 139À149. Nadeau P.H., Wilson M.J., McHardy W.J. & Tait J. Pevear D.R., Dethier D.P. & Frank D. (1982) Clay (1984) Interstratified clays as fundamental particles. minerals in the 1980 deposits from Mount St Helens. Science, 225, 923À925. Clays Clay Miner. 30, 241À252. Nahon D.B. & Colin F. (1982) Chemical weathering of Pevear D.R., Goldin A. & Sprague J.W. (1984) Mineral orthopyroxenes under lateritic conditions. Am. J. Sci. transformations in soils formed in glacial marine 282, 1232À1243. drift, Northwestern Washington. Soil Sci. Soc. Am. J. Nash V.E., Pettry D.E. & Mohd Noor Sudin (1988) 48, 208À216. Mineralogy and chemical properties of two Ultisols Poncelet G.M. & Brindley G.W. (1967) Experimental formed in glauconitic sediments. Soil Sci. 145, formation of kaolinite from montmorillonite at low 270À277. temperatures. Am. Miner. 52, 1161À1173. Nettleton W.D., Nelson R.E. & Flach K.W. (1973) Proust D. (1982) Supergene alteration of metamorphic Formation of mica in surface horizons of dryland chlorite in an amphibolite from the Massif Central, soils. Soil Sci. Soc. Am. Proc. 37, 473À478. France. Clay Miner. 17, 159À173. Nettleton W.D. & Peterson F.F. (1983) Aridisols. Pp. Proust D., Eyinery J.P. & Beaufort D. (1986) Supergene 24 M. J. Wilson

vermiculitization of a chlorite: iron and Weathering of appinite and Ivy Scar rock. J. Soil Sci. magnesium removal processes. Clays Clay Miner. 3, 219À237. 34, 572À580. Sudo T. & Yotsumoto H. (1977) The formation of Rausel Colom J.A., Sweatman T.R., Wells C.B. & halloysite tubes from spherulitic halloysite. Clays Norrish K. (1965) Studies in the artificial weathering Clay Miner. 25, 155À159. of mica. Pp. 40À72 in: Experimental Pedology (E.G. Tardy Y., Cheverry C. & Fritz B. (1974) Neoformation Hallsworth & D.V. Crawford, editors). Butterworths, d’une argile magnesienne dans les de´pressions London. interdunaires du lac Tchad. Applications aux do- Reid D.A., Graham R.C., Edinger S.B., Baurent B.H. & maines de stabilite´ des phyllosilicates alumineux, Ervin J.D. (1988) Celadonite and its transformation magnesiens et ferriferes. C. R. Acad. Sci. Ser. D. to smectite in an Entisol at Red Rock Canyon, Kern 278, 1999À2002. County, California. Clays Clay Miner. 36, 425À431. Tomura S., Shibasaki Y. & Mizuta H. (1985) Growth, Robert M., Hardy M. & Elsass F. (1991) conditions and genesis of spherical and platy Crystallochemistry, properties and organization of halloysite. Clays Clay Miner. 33, 200À206. soil clays derived from major sedimentary rocks in Trauth N., Paquet H., Lucas J. & Millot G. (1967) France. Clay Miner. 26, 409À420. Montmorillonite of lithomorphic Vertisols are ferri- Romero R., Robert M., Elsass F. & Garcia C. (1992) ferous: geochemical and sedimentological conse- Evidence by transmission electron microscopy of quences. C. R. Acad. Sci. Paris Ser. D. 264, weathering microsystems in soils developed from 1577À1579. crystalline rocks. Clay Miner. 27,21À34. Ugolini F.C., Dahlgren R., La Manna J., Muhn W. & Ross G.J. & Mortland M.M. (1966) A soil beidellite. Zachara J. (1991) Mineralogy and weathering Soil Sci. Soc. Am. Proc. 30, 337À343. processes in Recent and Holocene tephra deposits Ross G.J. & Kodama H. (1976) Experimental alteration of the Pacific Northwest, USA. Geoderma, 51, of chlorite into a regularly interstratified chlorite- 277À299. vermiculite by chemical oxidation. Clays Clay Vicente M.A., Razzaghe M. & Robert M. (1977) Miner. 24, 183À190. Formation of aluminium hydroxy vermiculite (inter- Rossignol J.P. (1983) Les Vertisols du nord de grade) and smectite from mica under acidic condi- l’Uruguay. Cah. ORSTOM se´rPe´dol. 20, 271À291. tions. Clay Miner. 12, 101À112. Rutherford G.K. & Debenham P.L. (1981) The miner- Vicente-Hernandez J., Vicente M.A., Robert M. & alogy of some silt and clay fractions from some soils Goodman B.A. (1983) Evolution des biotites en on the Faeroe Islands. Soil Sci. 132, 288À299. fonction des conditions d’oxydo-reduction du milieu. Sawhney B.L. (1989) Interstratification in Layer Clay Miner. 18, 267À275. Silicates. Pp. 789À828 in: Minerals in Soil Villiers de J.M. & Jackson M.L. (1967) Aluminous Environments (J.B. Dixon & S.B. Weed, editors). chlorite origin of pH dependent cation exchange Soil Sci. Soc. America, Madison, Wisconsin. variations. Soil Sci. Soc. Am. Proc. 31, 614À619. Shadfan H. (1983) Clay minerals and potassium status Wada K. & Kakuto Y. (1983) Integradient vermiculite- of some soils of Jordan. Geoderma, 31,41À56. kaolin mineral in a Korean Ultisol. Clays Clay Singer A. (1971) Clay minerals in the soils of the Miner. 31, 183À190. southern Golan Heights. Isr. J. Earth Sci. 20, Wada S.I. & Mizota C. (1982) Iron-rich halloysite with 105À115. crumpled lamellar morphology from Hokkaido, Singer A. (1984) Clay formation in saprolites of igneous Japan. Clays Clay Miner. 30, 315À317. rocks under semi arid to arid conditions, Negev, Walker G.F. (1949) The decomposition of biotite in the Southern Israel. Soil Sci. 137, 332À340. soil. Mineral. Mag. 28, 693À703. Singh B. & Gilkes R.J. (1991) A K-rich beidellite from a Watanabe T., Sowada Y., Russell J.D., McHardy W.J. & lateritic pallid zone in West Australia. Clay Miner. Wilson M.J. (1992) The conversion of montmorillo- 26, 233À244. nite to interstratified halloysite-smectite by weath- Singh B. & Gilkes R.J. (1993) Weathering of spodu- ering in the Omi acid clay deposit, Japan. Clay mene to smectite in a lateritic environment. Clays Miner. 27, 159À173. Clay Miner. 41, 624À630. Wilding L.P., Drees L.R., Smeck N.E. & Hall G.F. Smeck N.E., Wilding L.P. & Hollawaychuk N. (1968) (1971) Mineral and elemental composition of Genesis of argillic horizons in Celina and Morley Wisconsin age till deposits in west-central Ohio. soils of western Ohio. Soil Sci. Soc. Am. Proc. 32, Pp. 290À318 in: Till: A Symposium (R.P. 550À556. Goldthwait, editor). Ohio State University Press, Soil Survey Staff (1960) Soil Classification: A Columbus, Ohio. Comprehensive System. USDA. Wilke D.M. & Zech W. (1987) Mineralogies of silt and Stephen I. (1952) A study of rock weathering with clay fractions of twelve soil profiles in the Bolivian reference to the soils of the Malvern Hills. II. Andes (Callavaya Region). Geoderma, 39, 189À208. Origin and formation of clay minerals in soils 25

Wilson M.J. (1965) Weathered biotite from Strathdon, Wilson M.J. & Mitchell B.D. (1979) Comparative study Aberdeenshire. Nature Lond. 210, 1188À1189. of a Vertisol and an Entisol from the Blue Nile plains Wilson M.J. (1966) Clay mineralogy of some soils of Sudan. Egypt J. Soil Sci. 19, 207À220. derived from a biotite-rich quartz gabbro in the Wilson M.J. & Nadeau P.H. (1985) Interstratified clay Strathdon area, Aberdeenshire. Clay Miner. 7, minerals and weathering processes. In: The 91À100. Chemistry of Weathering (J.I. Drever, editor). D. Wilson M.J. (1970) A study of weathering in a soil Riedel, Dordrecht, The Netherlands. derived from a biotite- rock. I. Wilson M.J., Bain D.C. & Duthie D.M.L. (1984) The Weathering of biotite. Clay Miner. 8, 291À303. Wilson M.J. (1987) Soil smectites and related inter- soil clays of Great Britain. II. Scotland. Clay Miner. stratified minerals: recent developments. Proc. Int. 19, 709À735. Clay Conf., Denver, 167À173. Wilson M.J., Oyegoke C. & Fraser A.R. (1997) Wilson M.J. (1993) Pedologic factors influencing the Occurrence of trioctahedral clay mica in some distribution and properties of soil smectites. Trends Nigerian alluvial soils. 11th Int. Clay Conf. Agric. Sci. 1, 199À216. Ottawa, Canada. Abstracts. p. A83.